Facebook launched its Deepfake Detection Challenge at the Conference on Neural Information Processing’s Systems on December 11th in Vancouver.
The challenge is a collaborative initiative that aims to create new technologies that can detect deepfakes and manipulated media.
The social media giant is working with a number of leaders in the industry and members of academia to facilitate the challenge. Amazon and Microsoft both previously announced their decision support the challenge.
Participants in the challenge will receive a new and unique data set of more than 100,000 videos that are designed to aid research on deepfakes. The participants are then going to use the data set to develop better models to detect manipulated media.
Facebook also announced that it is partnering with Kaggle, which is a data science and machine learning community site. Kaggle is hosting the challenge leaderboard.
The models that the participants develop are going to be ranked for effectiveness. The first winner of the challenge will receive $500,000 USD (approximately $659,000 CAD) in prize money. There are also prizes for second, third, fourth and fifth places.
The challenge is going to run until March 2020.